DIY LEDs - The write-up

Status
Not open for further replies.
Questions:
1 - How many leds do I need for each module ?

Depends on a lot of factors. I'd estimate between 80 - 100 LEDs total, so maybe 36 on each module.

2 - 50% cool white and 50% royal blue, is it right ?

Depends on what color you like. If you're building three modules, you might want to get a bunch of each color and play with the numbers to get the exact look you want. Most people seem to like 50/50. It's about equivalent to a 10kk MH with strong actinic supplementation.

3 - Do I need lenses for them (like I said, the modules will be at about 10-15 inches)

At 10 - 15" up and a moderately deep tank, I'd probably go with wide optics - 60-80 degrees.

Something I don't understand: the color temperature for the XR-E Q5 is 6500K.
the Wavelength for the royal blue is 450nm
4 - what is the conversion nm to K ??
5 - in MH the color temperature is between 10000K and 20000K. How many % of power do I need to put on the cool white and the royal blue to have 10000K and the same for 15000K

Sadly it's hard to make the conversions you're trying to make. There is no real way to convert from a wavelength of light to a kelvin color. Wavelength is a pure physical measurement - when a blue LED is labeled as 450nm, that means it peaks strongly at that exact wavelength, and has virtually no output elsewhere. Meanwhile, the kelvin scale is used to approximately describe "whitish" light, i.e. light from full spectrum sources that cover a very large wavelength. But even at that task it fails pretty spectacularly. Above I mentioned a very rough approximation, you can use that as a guide.

Another consideration is that the published specs (both wavelength for the blues and kelvin for the cool whites) are very rough guides in the first place. The LEDs are also labeled with a color bin. For colored LEDs like the royal blue, the color bin indicates the donminant wavelength. For cool white LEDs, the color bin indicates a polygon on a slice of the the CIE 1931 XYZ color space, which is basically a 2-d way to approximate the color of full spectrum light. Meanwhile, the kelvin scale is one dimensional (linear). This adds yet another important dimension to the consideration, because two color bins on that CIE 1931 graph might be assigned the same "kelvin" color, but to your eye, they may look very different.

That said, in general:

    • 60/40 cool white/royal blue will look like a plain 10kk MH lamp.
    • 50/50 cool white/royal blue will look like a 10kk MH with actinic supplementation
    • 40/60 cool white/royal blue will look like a 15kk MH with actinic supplementation

    And I'll repeat my suggestion above - people building smaller fixtures should get two or three extra of each color, then play with the counts to get the color they want. People making large fixtures should probably experiment with sub-units or smaller test fixtures first to get the balance they want.

    In contrast, many people buy dimmable drivers with the intention of dimming one color or the other to fine-tune the look of the finished product. That's fine, but it's important to note that doing so also impacts the overall intensity of the fixture. An innocent twist of a knob on a dimmable driver can impact intensity by a very large margin, so it's something to consider. IMHO a "best practice" approach would be to build the fixtures with the balance you want in the first place, then use dimming to adjust afterwards if you want to tweak it, with considerations for overall intensity.
 
Sorry, there must be some confusion. I was trying to compare ratios of cool white and royal blue LEDs to MH and actinic supplementation. You are correct - with an LED rig, you do not need actinic supplementation in the form of fluorescent lamps.
 
Why do you look at the color temps? It is not important - what is important is spectrum that coincides with that consumed by photosynthetic systems. My opinion is that just blue+white is not enough - some UV (~390-400nm) LEDs should be added.
Actinic bulbs are not blue (like royal blue) - they have that purplish color (intens spectral band at 400-450 nm) - same that comes from UV LEDs and it appears that there is a peak in chlorophyll's absorbtion spectra at this wavelength (~390 nm ). So adding UV LEDs to the whole system will improve the "quality" of spectra making it more "full spectrum".
 
Just because the optics are "cheap", doesn't mean it's not doing a good job. Have you ever seen one in person? Perhaps taken it apart? What makes an Icecap or ATI reflector higher end? Price?
I never said the optics being used on these LEDs were "not doing a good job." You came up with that on your own. I did say they are "cheap" meaning inexpensive... and they are.

Obviously you haven't been reading the threads of some of the folks here on RC that have put a lot of effort into comparing the T5 reflectors. There are PAR meters to compare performance of the reflectors. The ATI reflectors are probably the best performing T5 reflector you can buy right now. I think IceCap comes close. I was only trying to imply that a T5 reflector could give a wide "flood light type optics" as Skeptic_07 desired for his new XP-G. He stated he wanted a "wider angle optic."

The optics are designed for these LEDs, and I would say doing a better job gathering/focusing the light where it needs to go when compared with the so called reflectors on the Reefbrights.

As I said above, I was throwing out an option for a wider angle reflector. Not trying to "focus the light". Also, why would you say they the optics are "doing a better job"? What about taking your own advice with the T5 reflector? "Have you ever seen one in person?" Or better yet, used one with LEDs and taken PAR readings?

It's analogous to those ricers supping up their Hondas with inferior aftermarket parts. It demonstrates a lack of understanding with the technology.

Hmm, at least I was trying to compare an LED optic to a T5 reflector. You are comparing them to cars.

I think what many people still don't get that LEDs output light in a very different manner. The way they are mounted in that reflector does almost nothing to harvest light that would otherwise be spilling out the sides at 180˚ angles.
-R

Those reefbrite LEDs are putting out a ton of PAR using what appears to be a T5 reflector. The LEDs in those reflectors are also underpowered to only 2W per LED. I think the reflector would have some great potential for the DIYers.
 
Last edited:
oh ok lol

sorry too about sometimes my bad english :)

No problem. Sometimes I have bad English and I've been doing it my whole life. :D

Why do you look at the color temps? It is not important - what is important is spectrum that coincides with that consumed by photosynthetic systems. My opinion is that just blue+white is not enough - some UV (~390-400nm) LEDs should be added.
Actinic bulbs are not blue (like royal blue) - they have that purplish color (intens spectral band at 400-450 nm) - same that comes from UV LEDs and it appears that there is a peak in chlorophyll's absorbtion spectra at this wavelength (~390 nm ). So adding UV LEDs to the whole system will improve the "quality" of spectra making it more "full spectrum".

MPC, read back a few dozen pages, and/or go look in the other big LED threads. We've gotten on the subject of "which mix of LEDs will provide the best growth for our corals" questions a few times, and the consensus from the experts (i.e. the marine biologists who have participated, not the hacks like me) has been that the mixes we're using are probably fine, and there's generally no need to "fill out" and hit the entire response curve for photosynthesis.

And at any rate, many people have gotten very good results with the typical builds, which backs up the theory with real (though anecdotal) results.
 
Obviously you haven't been reading the threads of some of the folks here on RC that have put a lot of effort into comparing the T5 reflectors.

I appreciate those efforts but it's very risky to extrapolate the results to an LED. A T5 lamp is a very distributed and completely omni-directional source of light. An LED is a point source of highly directional light. In general, they require very different approaches when you're trying to alter where the light produced ends up going.

There would certainly be nothing wrong with mounting LEDs in a T5 reflector, but it would be largely a waste of the design effort that went in to the T5 reflector, since the LED is already projecting in a narrower band than the typical T5 reflector creates, so there's going to be precious little actual "reflecting" going on - the light from the LED is simply not going to be hitting most of the surface of the reflector - the reflector will not be illuminated. Meanwhile, if you put a T5 lamp in that reflector, the entire reflector is illuminated, which means it actually has an opportunity to do something with the light produced by the lamp.

This is in opposition to a purpose-designed LED optic, which takes the directionality and default viewing angle of the LED into account.

Hope that makes sense. I don't think there's any real debate here, just perhaps a misinterpretation.
 
I have been following this thread on another site and the person who built this fixture has posted the Arduino code on google (wasn't sure if I could post the link) as well as instructions on building a led controller. It controls sunrise, sunset and moonlight function for about $60. Can this be applied to parts that could be purchased in the US? I am not good with electronics so I am not sure.

I am currently building this controller. You can find everything except the screen readily in the US. However if you are not in a hurry I would just order the screen from http://www.nuelectronics.com/estore/index.php?main_page=product_info&cPath=1&products_id=12 . It is only 18 dollars with shipping, but it takes about a week, and a half to get it. If you are not good with electronics you will not have any problems with this as the instructions are a real no brainer. If you wanted to program it your self the code he provides is a great start.
 
I want to thank you for all your help my ligts are ur and running tonite thank you
now my next question is
optics or no optics?
if yes or no why?
if yes what sizes?
thank you very much

it is a 90 gal tank 48 / 24 / 18
 
Last edited:
I have been following this thread on another site and the person who built this fixture has posted the Arduino code on google (wasn't sure if I could post the link) as well as instructions on building a led controller. It controls sunrise, sunset and moonlight function for about $60. Can this be applied to parts that could be purchased in the US? I am not good with electronics so I am not sure.

is this it

http://code.google.com/p/dangerduino/downloads/list
 
Thats it. Just be aware that the code is designed for buckpucks. If you are going to use meanwells you will have to modify the code slightly.

Any chance that you or someone else in this forum could give us the corrected code and the correct meanwell drivers to use for this? I'm really interested in building this controller when I put my LED rig together.

Thanks,

Scott
 
Buckpucks have an inverted dim circuit - "ON" is translated to "OFF" and vice versa. That means you'd have to invert the PWM output of the Arduino to control a driver that didn't have an inverted input (i.e. meanwells, or many other drivers on the market. Buckpucks are kind of the odd one out in this respect). That's probably a simple code change, but I haven't looked at that code in detail to know exactly how you'd do it.

Also note that in this recent thread:

http://www.reefcentral.com/forums/showthread.php?t=1793374

it was noted that the mean well P drivers require a 10v PWM signal, so you'd likely have to rework that hardware, too.
 
Buckpucks have an inverted dim circuit - "ON" is translated to "OFF" and vice versa. That means you'd have to invert the PWM output of the Arduino to control a driver that didn't have an inverted input (i.e. meanwells, or many other drivers on the market. Buckpucks are kind of the odd one out in this respect). That's probably a simple code change, but I haven't looked at that code in detail to know exactly how you'd do it.

Also note that in this recent thread:

http://www.reefcentral.com/forums/showthread.php?t=1793374

it was noted that the mean well P drivers require a 10v PWM signal, so you'd likely have to rework that hardware, too.

you are correct as usual. To change th code is easy it is just time consuming, as there are many lines reflecting pwm value. As far as the voltage, yes you would need a separate board for the voltage difference not sure which one as I am doing this mobile atm. I believe it would be a 7227. Unfortunatly for the op the origiantor does not like meanwells, and has no plans to support them at this time. It seems as though he has shifted to I want to sell my sh!t mode. He is no longer offering support, instead he is trying to get sponsers for his product. Kind of messed up considering he is using an open source platform he paid nothing for. It's people like that holding back this hobby\obsession. When I get home I will post more on the changes, and as soon as I can finish my revised code, it will be posted, with my full (albeit limited) support. fwi you can controll either dimmable meanwell with this setup. I will try to compile both codes, and parts list at a later date. I am very excited about this project. It's why we use leds in the rirst place. Because of the infinate controllability. Isn't it?. This thread and it's "experts" has helped me, it's time to give back. Hope I can help.
 
Buckpucks have an inverted dim circuit - "ON" is translated to "OFF" and vice versa. That means you'd have to invert the PWM output of the Arduino to control a driver that didn't have an inverted input (i.e. meanwells, or many other drivers on the market. Buckpucks are kind of the odd one out in this respect). That's probably a simple code change, but I haven't looked at that code in detail to know exactly how you'd do it.

Also note that in this recent thread:

http://www.reefcentral.com/forums/showthread.php?t=1793374

it was noted that the mean well P drivers require a 10v PWM signal, so you'd likely have to rework that hardware, too.

you are correct as usual. To change th code is easy it is just time consuming, as there are many lines reflecting pwm value. As far as the voltage, yes you would need a separate board for the voltage difference not sure which one as I am doing this mobile atm. I believe it would be a 7227. Unfortunatly for the op the origiantor does not like meanwells, and has no plans to support them at this time. It seems as though he has shifted to I want to sell my sh!t mode. He is no longer offering support, instead he is trying to get sponsers for his product. Kind of messed up considering he is using an open source platform he paid nothing for. It's people like that holding back this hobby\obsession. When I get home I will post more on the changes, and as soon as I can finish my revised code, it will be posted, with my full (albeit limited) support. fwi you can controll either dimmable meanwell with this setup. I will try to compile both codes, and parts list at a later date. I am very excited about this project. It's why we use leds in the rirst place. Because of the infinate controllability. Isn't it?. This thread and it's "experts" has helped me, it's time to give back. Hope I can help.
 
I want to thank you for all your help my ligts are ur and running tonite thank you
now my next question is
optics or no optics?
if yes or no why?
if yes what sizes?
thank you very much

it is a 90 gal tank 48 / 24 / 18

Sorry I've lost track - how high up is your fixture? What's your LED count?


you are correct as usual. To change th code is easy it is just time consuming, as there are many lines reflecting pwm value.

That strikes me as poor code structure. Ideally, I'd assume you'd have a single function that did the PWM mapping. Then you could have lots of code elsewhere to determine what "value" the pin should be set at, but let that single function actually map your range of values to the 0-255 range the Arduino can produce. Maybe it's just personal preference though.

As far as the voltage, yes you would need a separate board for the voltage difference not sure which one as I am doing this mobile atm. I believe it would be a 7227.

You'd need a 10v source, not just a different transistor. FWIW, check the thread I posted above. I designed a shield that has a 10v source and transistors to convert the Arduino's PWM to a 10v signal. Waiting to get prototypes back from the board house, at which point I'll donate them to volunteers for testing since I don't have any mean wells to test with. I appreciate the all-in-one design (putting the drive circuits and LCD into a single package) in that circuit but it locks you in to those choices. Meanwhile if you break out the drive circuits onto a shield, you can just add it to any Arduino controller project - I like that approach because the circuitry required will vary depending on which LED driver you're using. Meanwhile, other reef-controller circuitry won't likely vary from tank to tank, so that stuff can be integrated without loss of flexibility.

Different approaches with different goals, though. :)
 
Sorry I've lost track - how high up is your fixture? What's your LED count?




That strikes me as poor code structure. Ideally, I'd assume you'd have a single function that did the PWM mapping. Then you could have lots of code elsewhere to determine what "value" the pin should be set at, but let that single function actually map your range of values to the 0-255 range the Arduino can produce. Maybe it's just personal preference though.



You'd need a 10v source, not just a different transistor. FWIW, check the thread I posted above. I designed a shield that has a 10v source and transistors to convert the Arduino's PWM to a 10v signal. Waiting to get prototypes back from the board house, at which point I'll donate them to volunteers for testing since I don't have any mean wells to test with. I appreciate the all-in-one design (putting the drive circuits and LCD into a single package) in that circuit but it locks you in to those choices. Meanwhile if you break out the drive circuits onto a shield, you can just add it to any Arduino controller project - I like that approach because the circuitry required will vary depending on which LED driver you're using. Meanwhile, other reef-controller circuitry won't likely vary from tank to tank, so that stuff can be integrated without loss of flexibility.

Different approaches with different goals, though. :)

This is true. I was only building this controller, as a proof of concept any way. You're idea is a much more feasible solution. I was trying to ultimately design a modular system, that could be as simple as swapping a shield to control the three "normal" ways to power our systems. I am relatively new to Arduino programming, so I am sure I could be more efficient. Your thread on this subject is a better place for this conversation. Thanks for the link. It has steared me in a better direction.
 
I didn't take it that way at all. Just trying to keep this thread on track. I is long and confusing enough without being hijacked. lol
 
Status
Not open for further replies.
Back
Top